The sequencing of individual human genomes may soon be routine in certain clinical contexts - for example, to diagnose suspected Mendelian disorders in pediatric patients, or to guide therapeutic decisions in cancer treatment. However, even as its cost plummets to $1,000 or less, the value of a personal genome will remain highly constrained by the poor interpretability of individual genetic variants. For example, although BRCA1 and BRCA2 are clinically actionable when loss-of-function mutations are present, and although both genes have been sequenced in >50,000 patients over the past decade, the result returned to patients is often still variant of uncertain significance. This challenge will profoudly deepen as clinical sequencing accelerates and as the list of clinically actionable genes grows. To address this, we propose to develop a novel approach for experimentally measuring the functional consequences of such variants of uncertain significance at an unprecedented scale, as well as innovative computational approaches for estimating the relative pathogenicity of any possible variant in the entire human genome. For clinically relevant genes, we will exploit massively parallel technologies for nucleic acid synthesis and sequencing towards a new paradigm for dissecting function at saturating resolution. The application of this paradigm will yield experimentally grounded predictions for the functional consequences of all possible single residue variants, thereby informing the interpretation of variants newly observed in patients. For the remainder of the human genome, we will develop a framework for integrating a proliferating diversity of coding and non-coding annotations to a single metric. We will then calculate this metric of relative pathogenicity for all possible single nucleotide variants in the human genome. We anticipate that these methods and the resulting pre-computations of pathogenicity will broadly enable the interpretation of human genome sequences in diverse clinical and research settings.